AI Paper Atlas

Research map / Learning theory and algorithms / Matrix and tensor methods

Power Variable Projection for Initialization-Free Large-Scale Bundle Adjustment

Simon Weber, Je Hyeong Hong, Daniel Cremers

ECCV 2024 · Learning theory and algorithms › Matrix and tensor methods · approach: Randomized sketching for matrices · 9 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use PoVar (power series with VarPro) instead of PoBA for the first stageinitialization-free BA, first stage · BAL (97 problems), tolerance 0.001performance profile (runtime and accuracy)Improved
use PoVar instead of PoBA at loose toleranceinitialization-free BA, first stage · BAL (97 problems), tolerance 0.01performance profileNo clear effect
use power series expansion instead of direct sparse Cholesky factorization (pOSE)first stage solver · BAL (97 problems)performance profileImproved
use expansion methods instead of preconditioned conjugate gradients (iterative)first stage solver · BAL (97 problems)runtime to reach cost thresholdImproved
use RiPoBA instead of RiPCG for the second stageRiemannian manifold second stage · BAL (97 problems), all tolerancesperformance profileImproved
combine PoVar (first stage) with RiPoBA (second stage)stratified BA pipeline · BAL (97 problems)performance profile (runtime and accuracy)Improved
vary pOSE coefficient eta (0.2, 0.3, 0.4 versus 0.1) (ablation)PoVar · BAL (97 randomly resampled problems per eta)performance profileNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy